How and Why to Use Experimental Data to Evaluate Methods for Observational Causal Inference
This work addresses the problem of limited empirical evaluation for causal inference methods in fields like medicine and economics, though it is incremental as it builds on existing RCT data to improve assessment.
The authors tackled the challenge of empirically evaluating causal inference methods by proposing Observational Sampling from Randomized Controlled Trials (OSRCT), which creates observational datasets with known treatment effects from RCTs, enabling a large-scale evaluation of seven methods across 37 datasets and revealing significant performance variations across data sources.
Methods that infer causal dependence from observational data are central to many areas of science, including medicine, economics, and the social sciences. A variety of theoretical properties of these methods have been proven, but empirical evaluation remains a challenge, largely due to the lack of observational data sets for which treatment effect is known. We describe and analyze observational sampling from randomized controlled trials (OSRCT), a method for evaluating causal inference methods using data from randomized controlled trials (RCTs). This method can be used to create constructed observational data sets with corresponding unbiased estimates of treatment effect, substantially increasing the number of data sets available for empirical evaluation of causal inference methods. We show that, in expectation, OSRCT creates data sets that are equivalent to those produced by randomly sampling from empirical data sets in which all potential outcomes are available. We then perform a large-scale evaluation of seven causal inference methods over 37 data sets, drawn from RCTs, as well as simulators, real-world computational systems, and observational data sets augmented with a synthetic response variable. We find notable performance differences when comparing across data from different sources, demonstrating the importance of using data from a variety of sources when evaluating any causal inference method.